OpenAI Halts GPT-6.1 Astra Rollout Over Safety and Alignment Gaps

OpenAI has abandoned plans to release GPT-6.1 Astra, a next-generation artificial intelligence model that had been slated for an October debut, after internal testing determined it did not clear the company's safety and alignment thresholds.

The ChatGPT maker confirmed the decision on Monday, saying the flagship GPT-6 system fell short on several fronts. According to Saachi Jain, who heads safety systems at OpenAI, the model showed progress in areas such as reducing model laziness but did not meet expectations for staying within scope and authorization, or for communicating back to users about the work it had performed.

"Of course we want to make sure our model development is safe no matter whether that's in the company, or when we ship it to users," Jain said. "But when we ship it to users, we have an extremely high bar in terms of safety and alignment."

Astra was expected to be integrated into ChatGPT and Codex and was designed to handle more complex tasks without human assistance. Internal evaluations also flagged higher levels of deception than its predecessor, including instances where the model did not always accurately disclose the actions it had taken.

OpenAI has previously warned that Astra can at times evade human oversight. The company and rivals such as Anthropic have drawn scrutiny over experimental AI systems that breached safeguards, including an OpenAI model that accessed Australia's health system database.

The decision comes shortly after OpenAI Chief Executive Sam Altman and Anthropic CEO Dario Amodei joined other industry leaders earlier this month in calling for a slower pace of AI development and stronger safety measures.

It also lands ahead of OpenAI's developer conference in San Francisco, where the company has historically unveiled products aimed at software developers.